Frequently Asked Questions (FAQ)

Yes, you may notice minor variations in the score. This happens due to the application's scoring architecture and the stochastic nature of AI (LLM) models.

• LLM-based Score Generation: The trust score is evaluated comprehensively by the Gemini AI rather than calculated via a rigid formula. As generative AI models are stochastic, they may produce slightly different scores for the exact same context.
• Search Query Generation Variability: English keywords used to search the medical literature database are dynamically generated by the AI. Different keywords yield slightly different reference results, leading to fluctuations in the final outcome.
• Cross-Validation Verdict Variability: For borderline claims where clinical evidence is not absolutely conclusive, the AI's detailed verdicts might shift slightly.

Do I need to run multiple checks then?

No, it is not necessary to run the check multiple times.
Although the score might shift, a claim rarely flips completely from "contradicted" to "verified". Minor adjustments usually remain within the same safety threshold. A single result provides a sufficiently reliable assessment of the overall medical context and associated risks.

This is highly likely due to a temporary API Rate Limit being exceeded, which activates the application's 'Demo (Mock) Mode'.

• Medifact uses the Google Gemini API to extract precise medical claims and generate comprehensive summaries.
• If the free API quota is exceeded due to high traffic, the system activates a safety fallback mechanism to prevent the entire service from crashing.
• When this fallback is active, instead of the real AI model, built-in mock data is used to simulate the process. This may result in pre-defined templates or generic sentences that might seem awkward or out of context.

How can I fix this?

Please rest assured that this is not a system error. Simply wait about a minute for the API quota to reset, and try your request again. You will then receive the normal, high-quality fact-check results analyzed directly by the Gemini AI.

In short, they are the same process. The "Start for Free" button on the Free Plan links to the standard sign-up page — there is no separate "Free Plan application" step. In other words, signing up is exactly what starts your Free Plan.

What you get on sign-up (Free Plan)

• 10 free credits: granted instantly, letting you run 10 fact-checks (1 credit each) for free.
• Permanent history: your verification results are saved permanently to your account.
• Full evidence sources: cross-referencing across PubMed, Cochrane, KCI, FDA, and MFDS is identical for free and paid users.

How it differs from guest use

• Guest (no login): up to 3 free trials without signing up, but records are session-based and not stored permanently.
• Member (Free Plan): 10 credits on sign-up and permanent history.
• Paid top-up: once credits run out, recharge via Basic ($9.9 / 100 credits) or Premium ($29 monthly / 500 credits + clinician report editing).

In summary, "sign-up" and "Start for Free (Free Plan)" are two labels for the same sign-up function, layered above the login-free 3-time guest trial and the paid top-up after credits are used.

Zero-Data Retention (ZDR) is a premium security feature where raw texts and analyzed fact-check results requested by a B2B partner are never written to our database or logs. They are completely and permanently erased from volatile memory (RAM) immediately after the response is sent.

• Sensitive Data Protection: Vital for hospitals, insurers, and pharmaceutical firms handling highly confidential medical records to comply with GDPR, HIPAA, and domestic privacy laws.
• Zero Leak Risk: Because no trace of raw information remains on our servers, there is a 0% threat of data exposure even in the event of an external breach.
• Gemini Enterprise Compliance: The data transit to Google Gemini API is bound to Enterprise-grade non-retention endpoints, ensuring the LLM provider does not retain or use your data for training.

Why is there a 10% surcharge?

It carries a 10% premium surcharge due to the added engineering complexity, including bypassing DB persistence, real-time memory pipeline maintenance, sanitizing system logs (masking PII), and additional security compliance audits.

The Dedicated Worker & SLA 99.9% Guarantee is a premium option designed to prevent B2B API requests from being throttled or delayed by surges in general public traffic. It provisions isolated background server threads (Workers) and legally commits to a 99.9% uptime Service Level Agreement (SLA).

• Dedicated Worker: Reserves private background processing queues solely for your traffic, bypassing public queues to ensure verification runs with zero wait time.
• SLA 99.9% Guarantee: Formally guarantees that monthly system downtime will not exceed 43 minutes. Contractual refunds or service credits are triggered if this standard is breached.
• Business Continuity: Crucial when the fact-check API is embedded directly in live consumer-facing systems (e.g., shopping checkouts or insurance portals) where a few seconds of downtime translates directly to lost revenue.

Why is there a monthly flat fee of $150?

Because maintaining dedicated high-priority queuing processes running 24/7, configuring multi-AZ redundant nodes, and dedicating active monitoring resources incurs constant server overhead regardless of your actual query count.